Digitally Assessed Nucleolar Size Compared to Traditional Pathologist WHO Grade Allocation in Clear Cell Renal Cell Carcinoma

Abstract

Traditional WHO grading for renal cell carcinoma (RCC), largely based on nucleolar prominence, has well-known intraobserver and interobserver variability. In the era of digital pathology, quantitative guidelines may be more accurate and reproducible. Herein, we investigate nucleolar size in clear cell RCC using digital quantification methods compared to traditional pathologist WHO Grade allocation. We reviewed nephrectomy specimens containing clear cell RCC at a subspecialized academic center, initially diagnosed as WHO Grade 2 or 3. Digital scans were initially reviewed without knowledge of the originally assigned WHO Grade. Within a 0.25 mm² area of the visually assessed highest tumor grade in each tumor, nucleolar diameters were digitally measured. Of 100 tumors reviewed, 50 (50%) were originally diagnosed as WHO Grade 2 and 50 (50%) as WHO Grade 3. The mean nucleolar diameter was significantly different between tumors assigned as WHO Grade 2 versus Grade 3 (2.087 ± 0.256 SD µm vs 2.813 ± 0.408 SD µm, p < 0.001). However, scatterplot evaluation showed significant overlap in digitally assessed nucleolar size between the grades, with 54% WHO Grade 2 and 44% WHO Grade 3 tumors showing nucleolar diameter between 2.07 and 2.76 um. Clear cell RCC assigned as WHO Grade 2 and Grade 3 by traditional methods show significant overlap in digitally measured nucleolar size, suggesting accuracy for grading based on nucleolar size could be improved. Additional outcome-based studies to identify appropriate size cut-offs and potential training of artificial intelligence models to automate grading may provide improved risk stratification by grade for clear cell RCC.